It s a presidential election year. The papers are full

Transcription

1 Don t Worry About the Election Just watch the Fed. Scott B. Beyer, Gerald R. Jensen, and Robert R. Johnson SCOTT B. BEYER is an assistant professor of finance in the College of Business of Northern Illinois University in DeKalb, IL. sbeyer@niu.edu GERALD R. JENSEN is a professor of finance in the College of Business of Northern Illinois University. gjensen@niu.edu ROBERT R. JOHNSON is an executive vice president of the CFA Institute in Charlottesville, VA. robert.johnson@cfainstitute.org It s a presidential election year. The papers are full of articles speculating on the relationship between security market returns and the outcome of the election. Is one political party better for the stock market? What about the influence of political gridlock on security returns? When Congress and the White House are controlled by different parties, does the market react? Movements in security prices are also routinely linked to the actions of the Federal Reserve. Is the Fed chairman the second-most powerful person on earth? Many studies link the political landscape and monetary conditions to performance of the security markets. Oddly enough, politics and economic conditions have not been considered jointly in an examination of security returns. Our analysis considers two dimensions of the political landscape the party of the president and political gridlock along with Fed monetary policy in examining long-term security returns. The findings indicate that Fed policy dominates political considerations in determination of security returns. And, contrary to popular opinion, we find compelling evidence indicating that political gridlock is not beneficial to security market performance. PREVIOUS EVIDENCE There are strong arguments linking political and monetary conditions to the state of the macro economy (e.g. Friedman and Schwartz [1963], Alesina, Roubini, and Cohen [1997], and Drazen [2000]). Related research SUMMER 2004 THE JOURNAL OF PORTFOLIO MANAGEMENT 1

2 completes the link by showing a corresponding relationship between security returns and political and monetary conditions. Analysts scrutinize political and monetary conditions before making investment recommendations. Fiscal policy and monetary policy represent fundamental variables believed to impact general business conditions, and thus security returns. While the two policies are set by independent bodies, they are frequently applied in a coordinated manner to achieve an economic objective. It is surprising the two variables have not been considered jointly in examination of security return patterns. Many studies have investigated the short-term or announcement-period reaction of the security markets to presidential elections and Fed policy announcements (e.g., Niederhoffer, Gibbs, and Bullock [1976] and Smirlock and Yawitz [1985]). Such short-term studies are designed to capture the market s assessment of a change in political or monetary conditions. Our analysis focuses instead on the long-term implications of shifts in political and monetary conditions. The aim is to determine the relationship between security returns and the policies implemented by the political parties and the Fed. Long-term security return patterns are more likely to correspond with monetary and fiscal policies that are actually enacted. Some research shows that stock returns are higher during Democratic administrations than during Republican administrations. Santa-Clara and Valkanov [2003] report that excess stock returns (stock returns less the T- bill rate) are 9% per year higher during Democratic administrations. The excess return difference for small stocks is even greater, reaching 22% for the smallest size-decile. 1 Santa-Clara and Valkanov note that the higher returns earned during Democratic administrations cannot be explained as compensation for higher risk, because the volatility of returns is actually higher during Republican administrations. Finally, they show that the return differences persist even after applying several robustness checks and controlling for changes in business conditions a political cycle puzzle. While the stock market has generally prospered during Democratic administrations, Johnson, Chittenden, and Jensen [1999] show that bond returns are much better for all maturities of bonds during Republican administrations. Bond returns are over twice as high in Republican administrations than Democratic administrations. Political gridlock is another dimension of the political landscape frequently associated with economic conditions and security returns. The economic rationale is that fiscal policy intervention is more likely to occur in times of political harmony rather than in political gridlock. 2 Popular press articles suggest gridlock is beneficial for the security markets. Shell [2001] quotes Tom McManus, chief equity strategist at Banc of America Securities, as saying The stock market loves gridlock. Now that the Republicans no longer have the majority in both houses, gridlock is back. Before the 2000 election, Butler [2000] quoted Edward Yardeni, chief investment strategist at Deutsche Banc Securities: Gridlock has been very good for the stock market. Just one day before that, Ip [2000] reviewed the political influences on the stock market, noting that since 1982, the market has soared while government has been divided for all but two years. The influence of Fed monetary policy on the security markets has been the subject of considerable analysis. Because we focus on long-term security returns associated with Fed policy actions, we do not detail the many studies of the announcement effect of Fed policy changes. 3 Several studies identify systematic patterns in longterm security returns that are associated with prior changes in monetary policy. Specifically, stock returns in periods the Fed is following an expansive policy are shown to be better than returns in periods of restrictive Fed policy. See as examples Jensen, Mercer, and Johnson [1996], Patelis [1997], Thorbecke [1997], and Jensen, Johnson, and Mercer [1998]. The monetary policy-related return patterns identified in these studies exhibit two characteristics consistent with the political cycles puzzle: 1) the patterns are substantially stronger for small stocks, and 2) the patterns cannot be attributed to differences in volatility or business conditions proxies. Our analysis extends this research in several ways. We believe we are the first to consider the political landscape and monetary conditions jointly in an examination of security return patterns. Monetary and fiscal policies are often structured in tandem to produce a desired result in the economy. The empirical similarities between the political cycles puzzle and the monetary policy-related return patterns provide further motivation for a joint examination of the two effects. Second, while political gridlock has been debated in the popular press, the relationship between gridlock and security returns has not been subjected to rigorous academic study. Third, previous studies have focused almost exclusively on stock returns. We evaluate equity indexes, fixed-income indexes, and the rate of inflation. Finally, we examine the relationship between security returns and political and monetary conditions over 2 DON T WORRY ABOUT THE ELECTION SUMMER 2004

3 subperiods to determine the consistency of the relationship over time. DATA AND METHODOLOGY We evaluate monthly returns for two equity indexes, two fixed-income indexes, and an inflation index. The sample period varies, depending on the availability of the data, starting in 1926 (the beginning of the return data) or 1937 (the beginning of the monetary policy data). Qualitative classifications are used to establish binary variables associated with the three alternative effects that are considered in the analysis. The alternative binary classifications are: Republican versus Democratic president, the presence or absence of political gridlock, and an expansive or restrictive Federal Reserve monetary policy stance. The monthly returns are obtained from Ibbotson Associates EnCon Analyzer [2004]. The indexes include large company stocks, small company stocks, long-term corporate bonds, Treasury bills, and inflation. The congressional and presidential calendar data are obtained from the Congressional Directory. Gridlock is determined by the two-year congressional period from January to January, compared to the party of the president for a four-year January-to-January period. 4 Researchers have relied on two different types of indicators to identify changes in monetary conditions. Several use refined measures of monetary conditions that are designed to identify a change in the degree of monetary stringency (e.g., Thorbecke [1997]), while others rely on broad indicators designed to identify a fundamental shift in Fed policy (e.g., Jensen and Johnson [1995]). Both types of measures are shown to have a significant relationship with security returns, which supports the robustness of the monetary policy return patterns and suggests that either measure is appropriate. Because we evaluate the long-term returns associated with major shifts in political and monetary conditions, we choose a broad monetary policy indicator, as it identifies a fundamental shift in monetary policy, much as a change in the presidency represents a fundamental shift in political conditions. The monetary policy variable is measured as a binary variable based on directional changes in the Fed discount rate. Specifically, an initial discount rate reduction initiates an expansive policy, while the first rate increase in a series initiates a restrictive policy. This classification system has been shown to differentiate monetary conditions effectively. For example, Jensen, Mercer, and Johnson [1996] show that this approach defines periods with significantly different levels and rates of change in monetary and reserve aggregates. To avoid any look-ahead bias, we reclassify the monetary environment the month following the announced change in the discount rate. 5 We conduct two broad types of statistical analysis to examine the relationship between security returns and political and monetary conditions. First, we use standard difference-in-means tests across various classifications to examine the primary relationships between security returns and 1) the party of the president, 2) political gridlock, and 3) monetary conditions. The primary relationships examine the variable of interest without controlling for either of the other two qualitative variables. Second, for each index, we estimate regressions with return as the dependent variable and three sets of qualitative variables corresponding to 1) the party of the president, 2) the monetary policy stance of the Federal Reserve, and 3) the presence or absence of political gridlock. The regression analysis allows us to control for the other two factors in examination of the relationship between a particular variable and security returns. Since security returns are likely to be influenced by both fiscal and monetary policy, failure to control for the two omitted variables may result in misspecified models and increase the likelihood of biased findings. RESULTS Exhibit 1 presents the mean annualized returns for the five index series under Republican and Democratic administrations. The two subperiods are defined to accomplish two goals. First, the sample is split approximately in half. The end of the first subperiod represents the confluence of the end of a monetary policy period, the end of a congressional cycle, and the end of a presidential cycle. The monetary policy results start in 1937, so 1968 falls at approximately the midpoint of the sample. Second, the beginning of the second subperiod corresponds with the Fed s announced change in the administration of the discount window in 1968 (see Froyen [1975]). Panel A of Exhibit 1 shows initial support for the notion that equity returns are higher during Democratic administrations. Under Democratic presidents, the S&P 500 returned 406 basis points more than the return earned under Republican presidents (but a statistically insignificant difference). More noteworthy, the returns for small stocks are substantially higher, 26.7% compared to 7.3%, SUMMER 2004 THE JOURNAL OF PORTFOLIO MANAGEMENT 3

4 E XHIBIT 1 Annual Security Returns and Political Administrations Panel A: Full Period S&P 500 Small-Cap Corp. Bonds T-Bills Inflation Republican Adm. (n = 420) a 10.78% 7.25% 8.22% 5.05% 2.29% Democratic Adm. (n = 480) 15.24% 26.70% 3.99% 2.75% 3.81% (Democratic Republican) b Mean 4.06% 18.25% -3.94% -2.20% 1.49% P-value Panel B: Subperiod Republican Adm. (n = 180) 9.49% 3.31% 3.64% 2.41% -1.34% Democratic Adm. (n = 336) 14.74% 28.45% 3.59% 1.39% 3.25% (Democratic Republican) Mean 4.83% 24.40% -0.05% -1.00% 4.65% P-value Panel C: Subperiod Republican Adm. (n = 240) 11.76% 10.29% 11.77% 7.08% 5.09% Democratic Adm. (n = 144) 16.42% 22.71% 4.91% 5.99% 5.12% (Democratic Republican) Mean 4.21% 11.35% -6.20% -1.02% 0.03% P-value a Number of months in each administration during the sample period. b The difference-in-means test was performed on mean monthly returns. Therefore, the P-value indicates whether the mean monthly returns are different between Republican and Democratic administrations. Annual returns are reported for expositional purposes and are calculated by multiplying mean monthly returns by 12. under Democratic than Republican presidents, both economically and statistically significantly different. These results are consistent with the findings in Johnson, Chittenden, and Jensen [1999] and Santa-Clara and Valkanov [2003], and suggest that fiscal policy actions enacted by Democratic presidents have been more favorable for equity performance than the policies enacted in Republican administrations. The Exhibit 1 findings show that fixed-income securities, both long-term and short-term, experienced superior performance during Republican administrations, however. The return differences are highly economically and statistically significant, suggesting that policy actions during Republican administrations were more conducive to stable or declining interest rates. Surprisingly, the real returns (returns less inflation) earned by fixed-income investors were substantial during Republican administrations, but during Democratic administrations were trivial in the case of corporate bonds and negative for T-bills. The inflation index suggests that fiscal policy actions during Democratic administrations, relative to Republican administrations, have been more likely to spur inflation. Results for the two subperiods are generally similar across the two periods, and support the view that equity returns are higher under Democratic administrations, while bond returns are higher under Republican ones. The return differences appear to be less significant for the equity index in the later period, which suggests that the relative benefits associated with the policies of Democratic administrations have diminished in the more recent period. Exhibit 2 shows the mean annualized returns for the five indexes by political gridlock. Large-caps (the S&P 500 index) performed very similarly during each of the two environments, 13.57% and 12.78%. The small-cap index, however, performed much better when there was no gridlock, 11.44% versus 23.53%. While the return difference for the small-cap index is sizable, it is statistically insignificant. These findings are contrary to the popular belief that the equity market prefers political gridlock. The return differences for the fixed-income indexes suggest that the bond market prospers during periods of political gridlock. Perhaps the inability to enact significant legislation during such periods serves to maintain a stable or declining interest rate environment. Interestingly, while gridlock appears advantageous for the bond market, it also corresponds with periods of higher inflation. The subperiod results for gridlock in Panels B and 4 DON T WORRY ABOUT THE ELECTION SUMMER 2004

5 E XHIBIT 2 Annual Security Returns and Political Gridlock Panel A: Full Period S&P 500 Small-Cap Corp. Bonds T-Bills Inflation Political Gridlock (n = 408) a 13.57% 11.44% 8.78% 5.53% 4.10% No Gridlock (n = 492) 12.78% 23.53% 3.64% 2.42% 2.27% (Gridlock No Gridlock) b Mean 0.71% % 4.97% 3.04% 1.79% P-value Panel B: Subperiod Political Gridlock (n = 96) 11.58% 9.91% 1.22% 1.95% 2.82% No Gridlock (n = 420) 13.19% 21.30% 4.16% 1.69% 1.36% (Gridlock No Gridlock) Mean -1.44% -9.54% -2.84% 0.25% 1.45% P-value Panel C: Subperiod Political Gridlock (n = 312) 14.19% 11.91% 11.20% 6.65% 4.49% No Gridlock (n = 72) 10.47% 28.19% 0.66% 6.74% 7.75% (Gridlock No Gridlock) Mean 3.40% % 10.48% -0.09% -3.04% P-value a Number of months in each condition during the sample period. b The difference- in- means test was performed on mean monthly returns. Therefore, the P-value indicates whether the mean monthly returns are different between periods with gridlock and no gridlock. Annual returns are reported for expositional purposes and are calculated by multiplying mean monthly returns by 12. C are confounded by the fact that gridlock prevailed during a majority of the later period but was seldom seen in the earlier period. In the earlier period, 96 of 516 months (18.6%) represented gridlock, while it was present in 312 of the 384 months (81.3%) in the later period. The results, however, clearly run counter to the popular belief that political gridlock is beneficial for the stock market. There is no evidence to support the claim that political gridlock was beneficial for the equity market in either the earlier or the later period. In fact, during both subperiods, the returns to the small-cap index were substantially higher during periods of non-gridlock, although the difference in returns is statistically insignificant in both cases. The results for the corporate bond and inflation index reverse over the two subperiods, suggesting no systematic relationship between it and either bond returns or inflation. Finally, the unbalanced sample sizes and the vast difference in inflation across the two subperiods produce the unusual observation that gridlock has a highly significant relationship with T-bill returns in the overall period, but no significant relationship in either subperiod. Exhibit 3 presents the mean annualized returns for the five indexes under periods of expansive and restrictive Fed monetary policy. The difference in the equity returns under the two monetary policies is quite striking. The S&P 500 is 970 basis points higher (p-value = 0.02), and the U.S. small-cap index is 1,919 basis points higher (p-value = 0.00) during expansive periods as opposed to restrictive periods. Clearly, the equity markets have prospered when the Fed has maintained an expansive policy stance. Exhibit 3 also shows higher returns for bonds under expansive relative to restrictive monetary periods (but not statistically significantly different at traditional levels). Unlike the equity and long-term bond indexes, the T-bill rate and inflation index indicate that short-term interest rates and the inflation rate are significantly higher during restrictive monetary periods. This would be consistent with expectations, as the Fed is generally considered to follow a restrictive monetary policy when inflation is a more prominent investor concern. The return patterns for the equity and long-term SUMMER 2004 THE JOURNAL OF PORTFOLIO MANAGEMENT 5

6 E XHIBIT 3 Annual Security Returns and Monetary Policy Panel A: Full Period S&P 500 Small-Cap Corp. Bonds T-Bills Inflation Expansive Policy (n = 377) a 18.48% 28.26% 7.13% 3.49% 2.99% Restrictive Policy (n = 383) 8.10% 7.72% 4.48% 4.92% 5.04% (Expansive Restrictive) b Mean 9.67% 19.19% 2.55% -1.36% -1.96% P-value Panel B: Subperiod Expansive Policy (n = 175) 13.85% 30.19% 2.88% 1.04% 2.25% Restrictive Policy (n = 201) 11.86% 12.27% 2.08% 2.37% 3.51% (Expansive Restrictive) Mean 1.80% 16.13% 0.78% -1.30% -1.22% P-value Panel C: Subperiod Expansive Policy (n = 202) 22.63% 26.60% 10.95% 5.66% 3.63% Restrictive Policy (n = 182) 4.08% 2.90% 7.19% 7.80% 6.74% (Expansive Restrictive) Mean 17.88% 23.09% 3.53% -1.99% -2.93% P-value a Number of months in each policy stance during the sample period. b The difference- in- means test was performed on mean monthly returns. Therefore, the P-value indicates whether the mean monthly returns are different between periods of expansive and restrictive monetary policy. Annual returns are reported for expositional purposes and are calculated by multiplying mean monthly returns by 12. bond indexes are accentuated by an evaluation of the real returns (nominal return less the inflation rate). The S&P 500 produced real returns of only 3.06% (8.10% 5.04%) and the small-cap index 2.68% (7.72% 5.04%) during periods the Fed took a restrictive policy stance. During expansive policy periods, the corresponding real returns for the two indexes were an ample 15.49% and 25.27%. Interestingly, the real return to the corporate bond index during restrictive monetary periods was a negative 56 basis points, which indicates that on average corporate bond investments were losing propositions in these periods. These findings provide compelling evidence that capital market investors have achieved superior returns during Fed expansive policy periods. While variance in returns is not reported, expansive policy periods are also characterized as periods with lower volatility, which further supports the attractiveness of these periods. The returns reported in Panels B and C of Exhibit 3 support the time series consistency of the return patterns identified in Panel A. While the findings are similar across the two subperiods, there is considerable evidence suggesting that the patterns strengthened in the later period. Specifically, in the later period, the difference in returns is substantially greater for each index, particularly the equity indexes. This finding is consistent with the contention of Friedman and Friedman [1979] that Fed policy was generally viewed as an ineffective tool following the market collapse of the 1930s. According to Friedman and Friedman, Fed policy was neglected in the years after the Great Depression because it was viewed as a weak instrument with little influence on economic conditions. Consistent with this view is the observation that during the 13-year period from 1937 through 1949 the Fed changed the discount rate only four times. Over the next five decades, though, the discount rate was changed 112 times, an average of 2.24 times per year. Obviously, the Fed became much more proactive in monetary policy in the years after the 1940s. Exhibit 4 considers the joint influence of the political landscape and Fed monetary policy stance. If these factors are interdependent and relevant for security market performance, as many have argued, the analyses presented in Exhibits 1 through 3 are likely to produce misleading 6 DON T WORRY ABOUT THE ELECTION SUMMER 2004

7 E XHIBIT 4 Security Returns and the Combined Influence of Political and Monetary Conditions Panel A: Regression Results Full Period ( ) Index S&P 500 Small-Cap Corp. Bond T-Bill Intercept ** (0.02) * (0.59) * * Inflation Panel B: Subperiod 1 ( ) S&P 500 Small-Cap Corp. Bond T-Bill * ** (0.02) ** (0.04) * * Inflation Panel C: Subperiod 2 ( ) S&P 500 Small-Cap Corp. Bond T-Bill Inflation (0.49) (0.06) (0.95) * * P-values in parentheses. *Statistically significant at the 1% level. **Statistically significant at the 5% level. Presidential Administration - (0.69) - (0.78) (0.45) * (0.76) (0.28) (0.65) (0.65) (0.14) * (0.14) (0.62) (0.55) * * findings. That is, failure to consider instrumental variables can result in either an understatement or an overstatement of the true strength of a relationship between variables. To examine the relationships jointly, we estimate multiple regressions including as independent variables the qualitative variables described earlier. 6 The presidential administration variable takes a value of 1 for a Republican president and 0 for a Democratic president. The political gridlock variable takes a value of 1 if gridlock is present and 0 if there is no gridlock. Finally, the monetary policy variable is 1 if Fed policy is expansive and 0 if Fed policy is restrictive. Political Gridlock (0.59) (0.19) ** (0.04) * (0.78) (0.36) (0.26) - (0.39) (0.14) (0.23) (0.36) (0.27) (0.12) (0.17) * Monetary Policy ** (0.02) * (0.16) * - * (0.82) (0.19) (0.85) * (0.27) * * (0.48) - * - * The regression results reported in Panel A confirm the important role of monetary conditions for security returns. After controlling for the political landscape, monetary policy has a significant relationship with the two equity indexes and is also significantly related to T-bill returns and the inflation rate. The party of the president by contrast is significantly related to T-bill returns only, and political gridlock is significantly related to only corporate bond returns and T-bill returns. The results in Panels B and C show a strong time series consistency in the coefficients on the monetary policy variable, but a general lack of consistency with Adj. R SUMMER 2004 THE JOURNAL OF PORTFOLIO MANAGEMENT 7

8 respect to the political landscape variables. In general, the high incidence of sign reversals across the two subperiods suggests that past associations between security returns and the political landscape have probably been spurious. For example, the presidential administration variable suggests that inflation was significantly lower under Republican administrations in subperiod 1, but it was significantly lower under Democratic administrations in subperiod 2. Likewise, Republican administrations are associated with higher equity returns in subperiod 1 but lower returns in subperiod 2, although the administration variable is insignificant in all equity models. Finally, gridlock has a positive relationship with inflation in subperiod 1, but a significant negative relationship in subperiod 2. The time series consistency in the coefficients on the monetary policy variable supports the robustness of the monetary policy relationship. This evidence suggests that, unlike the political landscape, monetary policy has a systematic relationship with security returns. The results across the two subperiods, however, do support the argument that monetary conditions have become a more prominent consideration for investors in the more recent period. This observation is consistent with the passive monetary policy approach followed by the Fed subsequent to the Depression. Overall, the results of our regression analysis suggest that the security return patterns that researchers have attributed to changes in the political landscape are instead more likely the result of monetary policy developments. While securities markets are undoubtedly influenced by fiscal policy actions, there is little evidence of a systematic relationship between security returns and either political gridlock or the party of the president. The evidence is compelling that there is a systematic relationship between changes in monetary policy and long-term security returns. CONCLUSIONS We know monetary and fiscal policy are interdependent, and both variables have been linked to patterns in security returns. Our analysis examines their joint effects on long-term security returns. The evidence is contrary to the popular opinion that the stock market benefits from political gridlock. Equity returns are found to be generally invariant to gridlock; if anything, equity markets perform more poorly under gridlock. While earlier evidence suggests that equities, particularly small stocks, earn better returns during Democratic administrations, our evidence suggests the relationship is spurious. After controlling for monetary conditions and gridlock, equity returns are higher in early periods under Democratic presidents, but higher in later years under Republican administrations. The relationships between security returns and presidential administrations are statistically insignificant in both the early and the later period, though. Earlier evidence had suggested that fixed-income returns, both short- and long-term, are higher during Republican administrations. Our evidence identifies such a relationship for T-bills, but not long-term corporate bonds. After controlling for shifts in the political landscape, we find strong evidence that shifts in Fed monetary policy have a significant relationship with security returns. The return patterns are particularly prominent during the last three decades, which is consistent with a contention that monetary policy was little used in the years following the Depression. The monetary policy-related return patterns, consistent with previous evidence, suggest that equity markets prosper when the Fed maintains an expansive policy stance. Expansive policy periods are also associated with significantly lower inflation, which indicates that real returns to investors are especially attractive during expansive periods. Overall, the evidence is that equity investors have consistently benefited from an expansive rather than restrictive Fed monetary policy. There is no consistent evidence suggesting that shifts in the political landscape have been systematically related to security returns. We d say market participants should focus on the actions of the Federal Reserve when they consider investment, but treat election outcomes as a minor distraction. ENDNOTES 1 The superior performance of the stock market during Democratic administrations is also noted by Huang [1985], Stovall [1992], and Johnson, Chittenden, and Jensen [1999]. 2 See, for example, Coleman [1999] for the theoretical rationale. 3 Examples are Waud [1970], Pearce and Roley [1985], and Smirlock and Yawitz [1985]. They show that the stock market reacts positively to announcements of Fed easing and negatively to announcements of policy tightening. 4 This is consistent with the research in this area. See Johnson, Chittenden, and Jensen [1999] and Santa-Clara and Valkanov [2003]. 5 For a more thorough discussion of the monetary policy 8 DON T WORRY ABOUT THE ELECTION SUMMER 2004

9 measure, see Jensen, Johnson, and Mercer [2000]. 6 The results are adjusted for correlation in the error terms and heteroscedasticity using the Newey and West [1987] general covariance matrix to determine the estimators and calculate the standard errors. The results are robust with the unadjusted least squares estimates. Jensen, G., R. Johnson, and J. Mercer. The Role of Monetary Policy in Investment Management. Charlottesville, VA: The Research Foundation of CFA Institute and Blackwell Series in Finance, Jensen, G., J. Mercer, and R. Johnson. Business Conditions, Monetary Policy and Expected Security Returns. Journal of Financial Economics, vol. 40, no. 2 (February 1996), pp REFERENCES Alesina, A., N. Roubini, and G. Cohen. Political Cycles and the Macroeconomy. Cambridge: MIT Press, Butler, S. Can the Bull Still Run? U.S. News & World Report, November 6, 2000, p. 38. Coleman, J. Unified Government, Divided Government, and Party Responsiveness. American Political Science Review, vol. 93, no. 4 (December 1999), pp Drazen, A. Political Economy in Macroeconomics. Princeton: Princeton University Press, EnCorr Analyzer TM. Chicago: Ibbotson Associates, Friedman, M., and R.D. Friedman. The Anatomy of Crisis. The Journal of Portfolio Management, vol. 6, no. 1 (Fall 1979), pp Friedman, M., and A. Schwartz. A Monetary History of the United States: Princeton: Princeton University Press, Froyen, R. The Determinants of the Federal Reserve Discount Rate Policy. Southern Economic Journal, vol. 42, no. 2 (October 1975), pp Huang, R. Common Stock Returns and Presidential Elections. Financial Analysts Journal, vol. 41, no. 2 (March/April 1985), pp Ip, G. Election Outcome Probably Won t Affect Long-Term Performance of the Market. Wall Street Journal, November 5, Jensen, G., and R. Johnson. Discount Rate Changes and Security Returns in the U.S., Journal of Banking and Finance, vol. 19, no. 1 (April 1995), pp Jensen, G., R. Johnson, and J. Mercer. The Inconsistency of Small-Firm and Value Stock Premiums. The Journal of Portfolio Management, vol. 24, no. 2 (Winter 1998), pp Johnson, R., W. Chittenden, and G. Jensen. Presidential Politics, Stocks, Bonds, Bills, and Inflation. The Journal of Portfolio Management, vol. 26, no. 1 (Fall 1999), pp Newey, W.K., and K.D. West. A Simple, Positive Semidefinite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix. Econometrica, vol. 55, no. 3 (May 1987), pp Niederhoffer, V., S. Gibbs, and J. Bullock. Presidential Elections and the Stock Market. Financial Analysts Journal, March/ April 1970, pp Patelis, A.D. Stock Return Predictability and the Role of Monetary Policy. Journal of Finance, vol. 52, no. 5 (December 1997), pp Pearce, D., and V. Roley. Stock Prices and Economic News. Journal of Business, vol. 58, no. 1 (January 1985), pp Santa-Clara, P., and R. Valkanov. The Presidential Puzzle: Political Cycles and the Stock Market. Journal of Finance, vol. 58, no. 5 (October 2003), pp Shell, A. Gridlock and the Financial Markets. USA Today, May 24, Smirlock, M.J., and J.B. Yawitz. Asset Returns, Discount Rate Changes, and Market Efficiency. Journal of Finance, vol. 40, no. 4 (September 1985), pp Stovall, R.H. Forecasting Stock Market Performance via the Presidential Cycle. Financial Analysts Journal, vol. 48, no. 3 (May/June 1992), pp Thorbecke, W. On Stock Market Returns and Monetary Policy. Journal of Finance, vol. 52, no. 2 (June 1997), pp Waud, R.N. Public Interpretation of Federal Reserve Discount Rate Changes: Evidence on the Announcement Effect. Econometrica, vol. 38, no. 2 (March 1970), pp To order reprints of this article, please contact Ajani Malik at amalik@iijournals.com or SUMMER 2004 THE JOURNAL OF PORTFOLIO MANAGEMENT 9